What the RSI is
The relative strength index, or RSI, is a momentum oscillator introduced by the engineer and analyst J. Welles Wilder in the late 1970s. It turns recent price changes into a single number between 0 and 100. The number is higher when recent gains have been larger than recent losses, and lower when losses have dominated.
It is drawn in a panel under the price chart and is among the most frequently applied indicators on any charting tool, crypto ones included. Its popularity says little about whether it works; the aim of this article is to show what it calculates so you can judge for yourself.
The formula
The standard version uses 14 periods. The calculation runs in four steps.
- Over the lookback window, record each period’s change in closing price.
- Compute the average gain, the average of the up-moves, counting down periods as zero.
- Compute the average loss, the average size of the down-moves, counting up periods as zero and ignoring the minus sign.
- Form the ratio RS = average gain / average loss, then:
RSI = 100 - 100 / (1 + RS)
Worked example
These numbers are invented to make the arithmetic easy.
Suppose that over the window the average gain is 1.5 and the average loss is 1.0.
- RS = 1.5 / 1.0 = 1.5.
- 1 + RS = 2.5.
- 100 / 2.5 = 40.
- RSI = 100 - 40 = 60.
Now suppose the average gain is 1.0 and the average loss is 3.0.
- RS = 1 / 3, about 0.333.
- 1 + RS = 1.333.
- 100 / 1.333 = 75.
- RSI = 100 - 75 = 25.
Two limiting cases help. If there were no losses at all, the RS would be unbounded and the RSI would be 100. If gains and losses are equal, RS is 1 and RSI is 100 - 50 = 50. So 50 is the neutral midpoint.
The smoothing step
After the first 14 periods, Wilder’s method does not recompute a plain average each time. It updates the averages with a smoothing rule:
new average gain = (previous average gain x 13 + current gain) / 14
and the same for losses. For example, if the previous average gain was 1.4 and the current period’s gain is 2.8, the new average is (1.4 x 13 + 2.8) / 14 = (18.2 + 2.8) / 14 = 21 / 14 = 1.5. This gives recent changes a bit more weight while letting older ones fade. Because of this, the exact RSI values can differ slightly between platforms depending on how they start the series.
Reading the scale
By common convention:
- Readings above 70 are called “overbought”.
- Readings below 30 are called “oversold”.
- The 50 line is used as a rough dividing line between stronger and weaker momentum.
Some people shift the thresholds to 80 and 20 for volatile assets. As with other settings, these are habits and not proven values.
Common misreadings
“Overbought means price will fall.” The label means only that recent gains were large relative to recent losses. In a strong advance RSI can remain above 70 for many periods while price keeps climbing. Selling purely because RSI crossed 70 has no support in the formula.
“Oversold means price will bounce.” The same logic applies in reverse. A falling market can stay oversold for a long time, and price volatility in crypto often produces long runs of extreme readings.
“RSI is a reliable predictor.” It is a transformation of past prices. It contains nothing that is not already in the chart, and it lags like any calculation built from history.
“Divergence is a sure warning.” Analysts point to price making a new high while RSI makes a lower high. It is subjective, since it depends on which peaks you pick, and divergences often occur without a reversal.
“One timeframe is enough.” An RSI of 80 on a five-minute chart and an RSI of 45 on the daily chart can both be true. Which one matters depends on what you are trying to understand.
“14 is the correct setting.” It is Wilder’s original choice for the markets of his day. Shorter periods make RSI swing more wildly; longer periods make it smoother and slower. There is no known best value.
Why the formula behaves as it does
Because RSI is a ratio of averages, it is bounded, so it can never leave the 0 to 100 range. That makes extremes look dramatic, but a bounded scale also compresses information: once RSI is near 90, a much bigger rise in price may change the number only slightly. Also, since RSI is based on closing-to-closing changes, it ignores what happened inside each period.
In trending markets, readings tend to hold in one half of the scale, for example staying above 40 in a rising market, while in sideways markets RSI swings across the whole range. Distinguishing between those cases is exactly the difficulty.
How to use it responsibly
A reasonable use of RSI is descriptive: it summarises whether the recent past was dominated by gains or by losses. If you want to test a rule built on it, use data the rule was not designed on and include costs, since backtests of oscillator rules are prone to overfitting. Other momentum indicators such as MACD offer similar information in a different form and are easily confused with independent confirmation, though they share the same inputs.
Cryptocurrency prices are volatile, past performance does not predict future results, and people can lose all they invest. This article is educational and is not financial advice.